fix(descriptors): convert Series to list in semantic_similarity_scoring

[RK0297]

This PR complements #1889 by fixing the second call site where a pandas.Series was passed directly into model.encode() in semantic_similarity_scoring (src/evidently/descriptors/_context_relevance.py).

Newer sentence-transformers versions (5.6.1+) no longer coerce pandas.Series objects, causing a ValueError: Unsupported input type: Series.

Changes

• Convert question.data.fillna("") and context_rows[context_column].fillna("") to Python list via .tolist() before calling model.encode().
• Add unit test in ests/descriptors/test_context_relevance.py with mocked SentenceTransformer.

Relates to #1888.

[Copilot]

Pull request overview

This PR fixes a runtime incompatibility in semantic_similarity_scoring where pandas.Series inputs were passed directly into SentenceTransformer.encode(), which newer sentence-transformers versions reject. It also adds a regression test to ensure encoding inputs are converted to plain Python lists.

Changes:

• Convert question.data.fillna("") and exploded context values to list via .tolist() before calling model.encode().
• Add a unit test verifying both encode() call sites receive list arguments.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
src/evidently/descriptors/_context_relevance.py	Converts filled Series inputs to Python lists before SentenceTransformer.encode() to avoid Unsupported input type: Series.
tests/descriptors/test_context_relevance.py	Adds a regression test covering the two encode() call sites and asserting list conversion.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.